Micron FQ1 FY2026: A Historic Quarter, and the Right Time to Take the Gain

Results vs. Consensus

Quality of Beat

• Revenue: Pure pricing power. DRAM revenue rose 20% sequentially to a record $10.8B on bit shipments up only slightly and prices up 20%. When a memory company grows revenue double-digits on price alone, with bits roughly flat, it is the definitional sign of a supply-constrained market at the height of its pricing leverage.
• Margins: The 56.8% gross margin beat the guide by 530 bps and rose 11 points sequentially, driven by pricing, cost execution, and mix. Operating margin reached 47%. Every business unit posted record revenue and sharply higher margins, with Cloud Memory at a 66% gross margin. This is as good as a memory income statement has ever looked.
• EPS: The $0.82 beat was overwhelmingly operational, with a normalized 15.1% tax rate. The cash quality was pristine: $8.4B of operating cash flow, $3.9B of record free cash flow, $2.7B of debt repaid, and a return to net cash. There is no asterisk on the quality of this print.

Segment Performance

Revenue by Technology

Revenue by Business Unit (all records)

Every segment at peak profitability

The new segment disclosure, in its second quarter, shows the entire portfolio re-rating at once. Cloud Memory's gross margin reached 66%, Core Data Center's jumped nearly ten points to 51%, and even the consumer-facing Mobile & Client unit expanded 17 points to a 54% margin on pricing alone. Core Data Center grew 51% sequentially, the fastest in the company, as AI server demand pulled high-capacity DRAM and SSDs.

"We saw sequential revenue growth across all our business units... Cloud Memory Business Unit revenue was a record $5.3 billion... CMBU gross margins were 66%." — Mark Murphy, CFO

Assessment: The breadth confirms an industry at maximum tightness: when the lowest-margin segment is earning 45% and the highest is at 66%, there is no internal mix lever left to pull. That is wonderful for current earnings and, for a cyclical, a signal that the operating leverage is largely spent. From here, incremental margin gains require still-higher prices, which is precisely what eventually invites supply.

Key Topics & Management Commentary

Overall Management Tone: The most bullish posture in the company's history, and management said so directly, calling Micron's competitive position "the best in its history" and declaring "the best is yet to come." The forward framing was uniformly confident: HBM sold out and locked through 2026, supply short beyond 2026, the long-term TAM pulled forward. The only restraint was on capex discipline, where management repeatedly emphasized it will not over-build, which is itself the tell that the capacity question is now front of mind.

1. Calendar-2026 HBM Fully Locked

Micron has completed price and volume agreements for its entire calendar-2026 HBM supply, including HBM4. This converts the single largest 2026 revenue driver into contracted backlog, removing the pricing uncertainty that was the chief open question in prior quarters.

"We have completed agreements on price and volume for our entire calendar 2026 HBM supply, including Micron's industry-leading HBM4." — Sanjay Mehrotra, CEO

Assessment: This is genuinely de-risking and is the strongest argument against our downgrade: calendar-2026 HBM earnings are now largely locked, so the most visible part of next year is not speculative. It is the reason this is a Hold and not a more cautious rating. But locked 2026 visibility does not address calendar 2027, which is where a $20B capex build begins to land as supply.

2. The TAM Pulled Forward

Management raised the HBM TAM trajectory to a roughly 40% CAGR reaching about $100B by 2028, two years earlier than the prior 2030 view, and noted the 2028 HBM TAM would exceed the size of the entire 2024 DRAM market. AI demand, management argued, has structurally changed memory from a component to a strategic asset.

"We forecast an HBM TAM CAGR of approximately 40% through calendar 2028, from approximately $35 billion in 2025 to around $100 billion in 2028... this 2028 HBM TAM projection is larger than the size of the entire DRAM market in calendar 2024." — Sanjay Mehrotra, CEO

Assessment: The secular HBM demand story is real and accelerating, and it is why memory may stay tighter for longer than in prior cycles. The risk is the standard one with pulled-forward TAMs at a cycle peak: the demand is rarely the problem, the supply response to extraordinary pricing is. A 40% TAM CAGR is also a powerful magnet for industry capacity.

3. Supply Short of Demand "Beyond 2026"

Management stated that aggregate industry supply will remain substantially short of demand for the foreseeable future, that tightness persists through and beyond calendar 2026, and that Micron can currently meet only 50-66% of several key customers' demand. The HBM trade ratio (3:1 with DDR5, rising with each generation) compounds the squeeze on standard DRAM.

"In the medium term, we are only able to meet about 50% to two-thirds of our demand from several key customers." — Sanjay Mehrotra, CEO

Assessment: The supply shortfall is the bull case in one sentence and the cycle risk in the same breath. Unmet demand of this magnitude guarantees a massive industry capacity response, and history says the cure for high memory prices is high memory prices. The tightness is real today; the question the rating turns on is how the market prices a peak it can now see the far side of.

4. Capex Raised to ~$20B

Fiscal-2026 capex was raised to roughly $20B (from $18B), weighted to the second half, primarily for HBM and 1-gamma. Management is pulling in equipment orders and accelerating Idaho construction (first wafer output pulled into mid-2027), while insisting it will remain disciplined and keep supply in line with demand.

"Micron is going to remain disciplined on CapEx growth to support bit demand... the plan is roughly to double the brick-and-mortar construction CapEx [from '25 to '26], and we would expect '27 CapEx to be up." — Mark Murphy, CFO

Assessment: This is the crux of the downgrade. A roughly $20B fiscal-2026 capex bill, doubling construction spend, with fiscal-2027 guided higher still, is a classic late-cycle capacity surge. Management's repeated discipline language is reassuring, but the entire industry is responding to the same signal, and the supply those dollars create lands in 2027-2028, precisely when the locked-2026 visibility runs out.

5. Multiyear LTAs With Specific Commitments

Management is negotiating multiyear contracts with several key customers spanning DRAM and NAND, structured differently from prior LTAs with specific commitments and a stronger contract structure, stretching toward 2027 and in some reports 2028.

"These are multiyear contracts that we are in discussions with several of our key customers... they have specific commitments in them and a much stronger contract structure." — Sanjay Mehrotra, CEO

Assessment: Genuine LTAs with volume and pricing commitments would be a structural change that legitimately lengthens the cycle and partially answers the 2027 question. This is the development most likely to pull us back to Outperform if it firms up. For now it is in discussion, not signed, so it supports the Hold rather than overturning it.

6. Technology Leadership Extends

1-gamma is ramping as the primary DRAM bit-growth driver for 2026, with 1-delta and 1-epsilon in development; G9 NAND will become the largest NAND node later in fiscal 2026. Micron claims four consecutive DRAM technology-node leads and progressively faster yield ramps each node, plus a record calendar-2025 quality year.

Assessment: Process leadership is the durable moat that persists through the cycle and is the reason Micron should be owned on any reset. It is not a near-term rating driver at these valuations, but it is the foundation of why a future Outperform re-engagement is attractive.

7. Balance Sheet Reaches Net Cash

Record free cash flow of $3.9B funded $2.7B of debt reduction and a return to net cash (+$244M), alongside a $300M buyback permitted under the CHIPS agreement. Management expects to further strengthen the balance sheet through the year.

Assessment: A net-cash balance sheet entering a heavy capex year is exactly the position of strength you want a cyclical to hold at the top, and it materially de-risks the $20B build. It is also why a downgrade here is to Hold, not lower: the financial risk is minimal even if the cycle turns.

8. The AI Productivity Flywheel

Over 80% of Micron's professional workforce now uses generative AI (up tenfold year-on-year), with coding gains of 30%-plus and root-cause identification time cut roughly in half in cases. Micron frames AI as both a demand driver and an internal productivity engine.

Assessment: A real if hard-to-quantify cost and speed advantage that compounds over time. It reinforces the long-term competitive position without changing the near-term cyclical math that drives the rating.

Guidance & Outlook

The fiscal-Q2 guide is the largest single-quarter step-up in the company's history: revenue jumping 37% sequentially to $18.7B, gross margin to 68%, and EPS to $8.42. A single quarter's guided EPS now exceeds Micron's entire prior fiscal-year EPS. Management expects records in revenue, gross margin, EPS, and free cash flow for both fiscal Q2 and the full year, with performance strengthening through the year.

Implied trajectory: Annualizing the $8.42 fiscal-Q2 EPS yields a run rate north of $33. The market is not being asked to forecast a recovery; it is being handed a guided peak.

Street at: The guide is far above any prior consensus, and estimates will gap higher. The relevant question is not whether near-term numbers rise, they will, but whether a stock that already trades at a single-digit multiple of that guided run rate can produce 12-month upside once the market begins discounting the eventual normalization.

The tension: Everything operational points up. The valuation already reflects a market that assumes these earnings are not durable, which is exactly why the multiple is so low. That is the paradox a cyclical investor must respect.

Analyst Q&A Highlights

The structure and duration of the multiyear contracts

The opening line of questioning probed the new long-term agreements being discussed, including reports of DDR5 bundled with HBM and NAND, stretching into 2027 and beyond. Management confirmed multiyear DRAM-and-NAND contracts with specific commitments and a stronger structure than prior LTAs, but declined to give specifics.

Q: "I know we are hearing about DDR5 that's being bundled with HBM, and in some cases, even NAND... it sounds like these are stretching out through 2026 and in some cases even 2027... can you talk about the nature of these LTAs?"
— Timothy Arcuri, UBS

A: "These are multiyear contracts that we are in discussions with several of our key customers... they have specific commitments in them and a much stronger contract structure. And beyond that, I cannot be giving you specifics at this point."
— Sanjay Mehrotra, CEO

Assessment: If these LTAs sign with genuine multiyear volume and price commitments, they would be the single most important structural change to the memory cycle and the most likely catalyst to pull us back to Outperform. That they are still "in discussions," not signed, is why we wait.

The capex step-up and capital intensity

A question pressed on the roughly $20B net fiscal-2026 capex, the implied sub-35% capital intensity, and whether spending rolls into a higher fiscal-2027. Management confirmed capex is rising primarily for DRAM, HBM, and 1-gamma, that construction spend roughly doubles year-on-year, and that fiscal-2027 capex would be up, while stressing discipline.

Q: "You took it up to $20 billion net... it seems like it is like 25 to 30%, which is a little below your 35% metric... is that because you are constrained because of fab space? And does that sort of roll into fiscal '27 where we would see capex more near that 35% range?"
— Timothy Arcuri, UBS

A: "A substantial part of that CapEx is to support DRAM and specifically HBM and the 1-gamma... from '25 to '26, the plan is roughly to double the brick-and-mortar construction CapEx. And at this time, we would expect '27 CapEx to be up... Micron is going to remain disciplined on CapEx growth."
— Mark Murphy, CFO

Assessment: A doubling of construction capex with fiscal-2027 guided higher is the supply response that, industry-wide, eventually ends every memory up-cycle. Management's discipline is credible, but the capacity these dollars create lands in 2027-2028, the period the locked-2026 visibility does not cover.

Why not add clean room faster given the shortage?

A follow-up questioned the seemingly conservative capacity expansion against an acute shortage, asking whether Micron is being deliberately judicious. Management explained that clean room takes time, node transitions are the principal 2026 supply source, and the whole industry will be short, while it pulls in tools and accelerates Idaho.

Q: "The relative [capex] growth seems very conservative in the backdrop that we are in... it feels like you are just sitting here without clean room space... can you talk about the philosophy there?"
— C.J. Muse, Cantor Fitzgerald

A: "Clean room space takes time. And the HBM growth, which has only picked up with AI-driven demand, has further pressured supply. So there is no near-term solution... the entire industry we expect to be short to demand. And we are no different in that case."
— Mark Murphy, CFO

Assessment: The candid admission that there is "no near-term solution" to the shortage is bullish for 2026 pricing and is why the cycle has room to run near-term. It also underscores the urgency behind the capex surge, which is the thing that resolves the shortage on a 2027-2028 horizon, on the far side of where the rating looks.

Cost trajectory and the HBM3E-to-HBM4 transition

A question sought the 2026 cost-down outlook across DRAM and NAND and any margin impact from the HBM3E-to-HBM4 transition. Management cited good cost execution, some new-fab start-up costs landing in 2026-2027 (small at current scale), and 1-gamma and G9 ramps as cost tailwinds, with HBM4 expected to ramp faster than HBM3E.

Q: "As you go through calendar '26, how should we think about cost down across both DRAM and NAND? And as you transition from 3E to 4, is there anything... where there might be higher cost temporarily given yields or whatnot?"
— C.J. Muse, Cantor Fitzgerald

A: "Our cost execution has been very good across both DRAM and NAND... We do have some startup costs coming in for the new fabs... at these levels, it is a relatively small impact on margin... 1-gamma DRAM and G9 NAND... those ramps are proceeding well and will be a tailwind to our cost."
— Mark Murphy, CFO

Assessment: Cost is a tailwind and the HBM4 yield ramp is tracking ahead, so the margin story has no near-term operational cracks. The risk to margin is not cost, it is the eventual price normalization, which is a 2027 question that today's guidance cannot speak to.

Managing HBM3E upside from rising ASIC demand

A question highlighted upward revisions to ASIC accelerator volumes (Google TPU, AWS Trainium) that still use HBM3E, asking how Micron manages that near-term 3E upside alongside a strong HBM4 profile given it is fully contracted for 2026. Management framed both HBM3E and HBM4 as carrying strong profiles in 2026 revenue, with mix driven by customer demand.

Q: "There has been a significant upward revision on ASIC XPU volume shipments next year... all of these XPUs are still going to be using HBM3E... given that you are fully contracted for calendar '26... how is the Micron team going to manage this upside dynamic in 3E alongside a strong HBM4 demand profile?"
— Harlan Sur, JPMorgan

A: "Our mix of HBM3E and HBM4 during '26 will be very much based on our overall customer demand, and we will have both of these products with a strong profile in our '26 revenue."
— Sanjay Mehrotra, CEO

Assessment: Demand is so strong that the operational challenge is allocating scarce supply between two well-priced products, an enviable problem. It reinforces that 2026 is locked and bullish, and says nothing to soften the 2027 supply question on which the rating turns.

What They're NOT Saying

1. A full fiscal-2026 revenue guide: Management guided only fiscal Q2, declining a full-year number even while calling for full-year records. The absence makes capital intensity (and the true scale of the build) hard to pin down.
2. Calendar-2027 supply-demand: Every visibility statement stops at "through and beyond 2026" without specifics on 2027, which is precisely when the $20B capex build converts to supply. The silence on the post-peak year is the most important omission.
3. Signed LTAs, not "in discussions": The multiyear contracts that would most de-risk 2027 remain unsigned. Management would not put terms, duration, or counterparties on them.
4. Where HBM4 pricing lands versus HBM3E: Volume and price are "locked," but the direction and level of HBM4 pricing relative to HBM3E went undisclosed, leaving the 2026 margin mix partly opaque.
5. Any acknowledgment of cycle risk: Management offered no framing of what a normalization looks like or when supply additions across the industry might rebalance the market. At a guided 68% gross margin, the absence of any down-cycle discussion is itself notable.

Market Reaction

• Pre-print setup: MU closed at $225.52 on December 17, up a remarkable 168% year-to-date, though it had cooled about 7% over the trailing 30 days into the print, leaving some room to run.
• After-hours / open: The stock gapped up 13.8% to open at $256.53 on the blowout and the historic fiscal-Q2 guide.
• Reaction session: Shares closed +10.2% at $248.55 (a $23 gain) on December 18, on 65.0M shares (2.5x the 30-day average), pulling back from the open but holding a double-digit gain. The S&P 500 was +0.8%.

This was the first clearly positive reaction in the four quarters we have covered, breaking a three-quarter pattern where strong prints met muted or negative tapes. The difference was the fiscal-Q2 guide: a 68% gross margin and $8.42 EPS were too large for even an extended stock to ignore, and the prior 30-day cooling-off had relieved some of the positioning pressure. The fade from the +13.8% open to a +10.2% close is the detail worth noting: even on a historic guide, the stock could not hold its highs, a hint that the marginal buyer is beginning to look across the peak.

Street Perspective

Debate: Structural super-cycle or the best peak memory has ever printed?

Bull view: AI has structurally changed memory: HBM locked through 2026, a $100B TAM by 2028, multiyear LTAs with commitments, and supply short beyond 2026. This is a durable super-cycle, not a spike, and current earnings are a base, not a peak.

Bear view: A 68% gross margin and a $20B capex surge are the textbook definition of a memory peak. Every prior "structural" memory thesis ended the same way: extraordinary pricing pulled forward extraordinary supply, and the stock topped well before earnings did.

Our take: The structural elements (HBM trade ratio, LTAs, data-center majority) are real and likely lengthen this cycle versus prior ones, which is why we move to Hold rather than turn negative. But "longer" is not "permanent," and at peak-of-record margins the burden of proof is on durability. Until the LTAs sign and 2027 supply is visible, we treat these as peak-adjacent earnings.

Debate: Is a single-digit forward multiple cheap or a warning?

Bull view: At roughly $248 against a $33-plus annualized fiscal-Q2 EPS run rate, the stock trades under 8x forward earnings with a net-cash balance sheet. That is far too cheap for a company growing this fast with locked HBM demand.

Bear view: Hyper-cyclicals always look cheapest at the peak, because the market is already capitalizing the coming earnings decline. A single-digit multiple on peak EPS is not value, it is the market pricing mean reversion.

Our take: This is the central question, and it resolves in the bear's favor for rating purposes. The low multiple is the tell, not the opportunity: the market is signaling it does not believe $33 of annualized EPS is the run rate. Buying a peak-cycle multiple expansion requires believing earnings keep rising past a guided 68% margin, which is a bet we will not make at this price.

Debate: Does the capex surge end the tightness or extend the boom?

Bull view: The $20B build is disciplined and demand-driven, funded by record free cash flow and a net-cash balance sheet; it extends Micron's supply leadership into a multiyear AI build-out without straining the company.

Bear view: Industry-wide, the same record pricing is triggering the same capacity response everywhere, and that supply lands in 2027-2028. Capex peaks are the most reliable leading indicator of the next memory down-cycle.

Our take: Micron's own build is prudent and well-financed; our concern is the industry aggregate. When every supplier responds to 68% margins at once, the 2027-2028 supply additions are the risk the current price does not reflect. The capex line is the single most important thing we will track from here.

Model Update & Valuation Framework

Valuation framework: At the $248.55 reaction close, the fiscal-Q2 guide of $8.42 EPS annualizes north of $33, putting the stock under 8x that run rate, with net cash. On the bull framing, that is conspicuously cheap for a company with locked 2026 HBM and a $100B TAM ahead. On the bear framing, a sub-8x multiple on a guided-peak earnings stream is the market's own forecast that the earnings normalize, and paying it is a bet on multiple expansion at the top of a cycle. The honest read is that the stock is simultaneously cheap on this year and fairly-to-fully priced on a through-cycle basis, and which one matters depends entirely on cycle timing.

Where we land: We move to Hold. We are not negative on Micron, which is executing better than at any point in its history with a net-cash balance sheet and locked 2026 demand. We are negative on the 12-month risk/reward after a triple from our June upgrade, at a 68% gross-margin guide and a surging capex bill, where the dominant historical pattern is that memory stocks de-rate before peak earnings arrive. We would return to Outperform on (1) a meaningful pullback that resets the risk/reward, (2) signed multiyear LTAs that credibly extend the earnings base into 2027-2028, or (3) evidence that calendar-2027 industry supply stays disciplined despite the capex surge. Absent those, we take the gain.

Thesis Scorecard Post-Earnings

Overall: The thesis has, in a sense, won: every bull pillar is confirmed and both near-term bears are resolved. But a thesis fully realized at a cyclical peak is precisely when risk/reward inverts. The escalating capital-intensity-and-cycle-maturity pillar is now the dominant consideration, and it is the basis for stepping aside. Note that our specific fiscal-Q4 downgrade triggers (free-cash-flow deterioration, supply catching demand) did NOT fire, the opposite did, so this downgrade rests on cycle position and valuation, not fundamental deterioration.

Action: Downgrade to Hold from Outperform. Take the gain from the June upgrade. Re-engage on a pullback, on signed LTAs that extend the earnings base, or on evidence of 2027 industry supply discipline.

Bottom Line

Micron just printed the best quarter in its history and guided to a better one: record revenue, a 56.8% gross margin, record free cash flow, a return to net cash, and a fiscal-Q2 outlook of 68% gross margin and $8.42 of EPS, a single quarter that out-earns all of fiscal 2025. Calendar-2026 HBM is fully sold out on locked price and volume, the long-term TAM was pulled forward two years, and supply is short of demand beyond 2026. On the fundamentals, there is nothing to criticize.

We are downgrading to Hold anyway, and we want to be precise about why. This is not a call against the business, which has never been stronger, and it is not our fiscal-Q4 triggers firing: free cash flow did not deteriorate (it set a record) and supply did not catch demand (the shortage deepened). It is a call about where we sit in the cycle and what we have already made. From our June upgrade the stock has tripled. It now trades at a 68% gross-margin guide that exceeds any prior peak, against a $20B capex bill that is doubling construction spend and rising again into fiscal 2027. Memory has taught the same lesson every cycle: the stock tops before the earnings do, and the cure for record prices is the record capacity that record prices invite. At a sub-8x multiple the market is already telling you it does not believe these earnings are the run rate. We agree, and we would rather book the gain than pay for multiple expansion at the peak. We move to the sidelines with full respect for how good the business is, and a clear list of what would bring us back: a pullback, signed multiyear LTAs, or visible 2027 supply discipline. Until then, Hold.